Assuming the observation of a squark at the Large Hadron Collider, we investigate methods to access its flavour content and gain information on the underlying flavour structure of the theory. Based on simple observables, we construct a likelihood inference method to determine the top-flavour content of the observed particle. In addition, we employ a multivariate analysis in order to classify different flavour hypotheses. Both methods are discussed within a simplified model and the more general Minimal Supersymmetric Standard Model. We conclude that the likelihood inference provides an estimation of the top-flavour content if additional knowledge on the gaugino sector is available, while the multivariate analysis identifies different flavour patterns and can accommodate a more minimalistic set of observables.